dave.j wrote:
My experience is the drifting starts from when the recording or streaming is started...it may well happen from when the unit is powered on...but I've not tested for that!
How soon the issues start seems to be effected by the frame rate...my tests at 25fps took up to 1hr50mins for the clicking to start, 60fps started around 20 minutes...so the higher the frame rate, sooner the issues start.
It seems to need an audio signal coming through the unit to trigger the drifting, my guess is that it's some kind of processing/buffer issue with the audio...it reaches a point where the amount of audio data passing through the USB connection exceeds the amount that can be processed, it starts to cache the extra data (this is where the clicking starts, little overruns roughly once a second)...after struggling with the overflow/clicking for 15-30 minutes it 'crashes' (hence the static/distortion/delayed signal) for about 30 seconds, then resets and the clicking is gone, except the audio is now 1 second behind.
That's just my best guess at what's happening...whether that processing is done/struggling within the ATEM unit or in the computer I don't know. The fact it happens across a range of old/new/windows/mac machines & the ATEM is the common factor suggests it's something with the processing inside the ATEM when sending the signal out via USB?
On the question about embedded HDMI drifting, I don't know. I've tested the Mic1/2 inputs alone, and together with audio from HDMI inputs, but not the HDMI inputs alone.
The workaround for now is to stream direct from the ATEM, I've run test for 2hrs 15mins without issues with the stream...and pretty sure I was recording USB to SSD during that test too.
Hi everyone,
Hope you're all well!
As Dave J has suggested, 1hr 50 seems to be crucial, at 25fps anyway... makes sense that it would happen sooner at higher frame rates...
I can confirm for me that it happens from when the unit is powered on, it doesn't have to be streaming or recording etc....
(EDIT: thinking about this now, I guess the unit could be powered on but no USB connected for a while and I bet this wouldn't then be an issue... I bet it is from the moment the USB is connected and it's actually sending something on the USB video and audio etc IE, when I connect it to the laptop and open vMix, it is then, from that point on, sending a signal down the USB to the computer, so the buffer will start filling etc… so it’s not specifically from when it is “streaming or recording”, as at the time you might not be doing that, it’s more from when it is capturing and sending information down the USB…)
We had the unit powered on at a location recording, recording into vMix, and luckily noticed the audio had gone out of sync just before we hit record on the next thing they wanted to film, I then re-booted the ATEM during problem solving which obviously fixed it... at that time I had no idea what the issue had been, but now I know that it had been on for more than 2hrs at this point... and it was just by chance I hadn't had headphones on when the USB audio would have been crashing then resetting out of sync...
This issue being something to do with the audio USB buffer makes total sense, I agree!
I ran the following test yesterday which confirmed the issue:
1. MacBook Pro - 2x HDMI Outputs into ATEM 3/4 and 1x MiniJack Audio Output into Mic/Line 2
2. Atem Mini Pro - USB C to USB A
3. RazerBlade 6 Core 9th Gen 2.6GHz / 16GB RAM / 256GB SSS running vMix
I played videos back to back from YouTube on one of the outputs of the MacBook Pro, TED talks or similar etc so that I could see if they sync went out...
Test timeline went as follows:0:00:00 - Turned on the laptop, ATEM Mini Pro and MacBook Pro from fresh
0:02:00 - Set the Youtube video running on input 4 of the ATEM and the audio level in Mic/Line 2
0:05:00 - Pressed record within vMix
0:05:00 to 1:52:00 - Everything ran fine... no issues of video or audio
1:52:00 - Started to hear ticking / clicking noise in the background (monitoring audio while it was recording through the laptop speakers)
2:15:30 - Ticking noise had continued from 1:52 getting worse... but at this point started it really loosing it's sh*t, sounding robotic, static, interference etc Effectively crashing the audio coming from the ATEM Mini Pro as others have said...
2:17:00 - Audio corrected itself after a lot of interference / breaking up... but... now the audio is out of sync with the video by 1 or more seconds, ie, very obvious....
Here is a link to the file where you hear the audio crashing, it's from the 2:00:00 onwards mark, so you can already hear ticking / clicking when it starts... but then at 15:30 into the video (so 2:15:30 actual time) you can hear the crash... (I obviously don't own the rights to the video at all, it was just whatever played next from YouTube etc!)
https://1drv.ms/u/s!AovxfgeFQlF3gplf8rK ... Q?e=R6caa8I then ran another test... This ran for just over 3 hours, just to make sure I was well past the 2hrs 15min mark, which was absolutely fine with no issues. This was using an external soundcard (Presonus 44VSL @ 48K, highest buffer setting, just in case, but it would probably be ok with a lower buffer setting) to bring the audio into vMix... So this is the workaround for now... Use an external soundcard... as the video stays in sync thorough out, it's the audio that has the issues...It seems that the USB audio bus is the issue within the ATEM Mini Pro, which would also make sense if others have had success streaming directly from the ATEM Mini Pro itself... as this wouldn't need / use the "webcam USB interface" ie, the USB audio buffer...
I hope they fix this soon, but, for now, external soundcard is the way! Cheers,
Chris